Generally described, embodiments of the present invention provide the ability to generate a universal query result set from several different search index result sets by including identifications of items from the different search index result sets in an accurate manner. After a query has been submitted, search index result sets are received from several different search indexes, an allocation score for each search index is computed and a universal item score for the top-level item identified in each search index is computed. The method then combines the allocation score and the universal item score for the top level item for each search index result set and adds the item with the highest combined score to the universal query result set.
Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A computer-implemented method for generating a universal query result set, comprising: under control of one or more computer systems configured with executable instructions, receiving a query from a searching entity; submitting the query to a plurality of search indexes, each search index corresponding to a respective category of items, at least one of the plurality of search indexes utilizing at least one of a different ranking property, scale, function, or definition for ranking items relative to other search indexes; receiving one of a plurality of search index result sets from each of the plurality of search indexes in response to the query; determining a plurality of appropriateness scores each corresponding to one of the plurality of search indexes, each of the plurality of appropriateness scores indicating an appropriateness of the category of items corresponding to the respective search index with respect to the query and being based at least in part on historical queries similar to the query that were submitted to the respective search index; determining a universal item score for each of a plurality of items in the plurality of search index result sets at least in part by normalizing the at least one of the different ranking property, scale, function, and definition to a ranking scale common to all the search index result sets; for each of the plurality of items, determining a probability that the item satisfies the query based at least in part on the appropriateness score for the search index associated with the item and the universal item score for the item; including in the universal query result set ones of the plurality of items selected in an order based at least in part on the probabilities of the plurality of items satisfying the query; and providing the universal query result set to the searching entity, wherein determining the plurality of appropriateness scores comprises modifying the plurality of appropriateness scores differently based at least in part on different types of recorded actions associated with the historical queries that were submitted to a corresponding search index.
A computer system creates a unified search result from multiple search indexes (e.g., web, products, images). Each index might rank results differently. The system receives a search query and sends it to each index. For each index, it calculates an "appropriateness score" based on how well that index's category suits the query, using past similar queries. It normalizes each item's rank from each index to a common scale to produce a "universal item score." For each item, it calculates the probability that it satisfies the query, based on the index's appropriateness score and the item's universal score. Results are shown in order of these probabilities. Appropriateness scores are adjusted based on past user actions (e.g., viewing, purchasing) for similar queries in that index.
2. A computer-implemented method according to claim 1 , wherein the method further comprises receiving the query from a searching entity.
A computer system creates a unified search result from multiple search indexes (e.g., web, products, images). Each index might rank results differently. The system receives a search query from a searching entity and sends it to each index. For each index, it calculates an "appropriateness score" based on how well that index's category suits the query, using past similar queries. It normalizes each item's rank from each index to a common scale to produce a "universal item score." For each item, it calculates the probability that it satisfies the query, based on the index's appropriateness score and the item's universal score. Results are shown in order of these probabilities. Appropriateness scores are adjusted based on past user actions (e.g., viewing, purchasing) for similar queries in that index.
3. A computer-implemented method according to claim 1 , wherein each of the plurality of search index result sets is ranked in accordance with a different internal relevance ranking function.
A computer system creates a unified search result from multiple search indexes (e.g., web, products, images). Each index might rank results differently. The system receives a search query and sends it to each index. For each index, it calculates an "appropriateness score" based on how well that index's category suits the query, using past similar queries. It normalizes each item's rank from each index to a common scale to produce a "universal item score." For each item, it calculates the probability that it satisfies the query, based on the index's appropriateness score and the item's universal score. Results are shown in order of these probabilities. Appropriateness scores are adjusted based on past user actions (e.g., viewing, purchasing) for similar queries in that index. Each search index result set uses a different internal relevance ranking function.
4. A computer-implemented method according to claim 1 , wherein at least one of the historical queries is associated with a recorded action of selecting a thereby identified item for viewing.
A computer system creates a unified search result from multiple search indexes (e.g., web, products, images). Each index might rank results differently. The system receives a search query and sends it to each index. For each index, it calculates an "appropriateness score" based on how well that index's category suits the query, using past similar queries. It normalizes each item's rank from each index to a common scale to produce a "universal item score." For each item, it calculates the probability that it satisfies the query, based on the index's appropriateness score and the item's universal score. Results are shown in order of these probabilities. Appropriateness scores are adjusted based on past user actions (e.g., viewing, purchasing) for similar queries in that index. Historical queries are associated with viewing an item.
5. A computer-implemented method according to claim 1 , wherein at least one of the historical queries is associated with a recorded action of selecting a thereby identified item for potential purchase.
A computer system creates a unified search result from multiple search indexes (e.g., web, products, images). Each index might rank results differently. The system receives a search query and sends it to each index. For each index, it calculates an "appropriateness score" based on how well that index's category suits the query, using past similar queries. It normalizes each item's rank from each index to a common scale to produce a "universal item score." For each item, it calculates the probability that it satisfies the query, based on the index's appropriateness score and the item's universal score. Results are shown in order of these probabilities. Appropriateness scores are adjusted based on past user actions (e.g., viewing, purchasing) for similar queries in that index. Historical queries are associated with selecting an item for potential purchase.
6. A computer-implemented method according to claim 1 , wherein at least one of the historical queries is associated with a recorded action of purchasing a thereby identified item.
A computer system creates a unified search result from multiple search indexes (e.g., web, products, images). Each index might rank results differently. The system receives a search query and sends it to each index. For each index, it calculates an "appropriateness score" based on how well that index's category suits the query, using past similar queries. It normalizes each item's rank from each index to a common scale to produce a "universal item score." For each item, it calculates the probability that it satisfies the query, based on the index's appropriateness score and the item's universal score. Results are shown in order of these probabilities. Appropriateness scores are adjusted based on past user actions (e.g., viewing, purchasing) for similar queries in that index. Historical queries are associated with purchasing an item.
7. A computer-implemented method according to claim 1 , wherein at least one of the historical queries is associated with a recorded action of selecting a category of a thereby identified item.
A computer system creates a unified search result from multiple search indexes (e.g., web, products, images). Each index might rank results differently. The system receives a search query and sends it to each index. For each index, it calculates an "appropriateness score" based on how well that index's category suits the query, using past similar queries. It normalizes each item's rank from each index to a common scale to produce a "universal item score." For each item, it calculates the probability that it satisfies the query, based on the index's appropriateness score and the item's universal score. Results are shown in order of these probabilities. Appropriateness scores are adjusted based on past user actions (e.g., viewing, purchasing) for similar queries in that index. Historical queries are associated with selecting a category of an item.
8. A computer-implemented method according to claim 1 , wherein determining the plurality of appropriateness scores comprises determining the plurality of appropriateness scores based at least in part on times associated with the historical queries that were submitted to a corresponding search index.
A computer system creates a unified search result from multiple search indexes (e.g., web, products, images). Each index might rank results differently. The system receives a search query and sends it to each index. For each index, it calculates an "appropriateness score" based on how well that index's category suits the query, using past similar queries. It normalizes each item's rank from each index to a common scale to produce a "universal item score." For each item, it calculates the probability that it satisfies the query, based on the index's appropriateness score and the item's universal score. Results are shown in order of these probabilities. Appropriateness scores are adjusted based on past user actions (e.g., viewing, purchasing) for similar queries in that index. Appropriateness scores are based on the times the historical queries were submitted.
9. A computer-implemented method according to claim 1 , wherein determining the plurality of appropriateness scores comprises determining the plurality of appropriateness scores based at least in part on how similar the historical queries submitted to a corresponding search index are to the query.
A computer system creates a unified search result from multiple search indexes (e.g., web, products, images). Each index might rank results differently. The system receives a search query and sends it to each index. For each index, it calculates an "appropriateness score" based on how well that index's category suits the query, using past similar queries. It normalizes each item's rank from each index to a common scale to produce a "universal item score." For each item, it calculates the probability that it satisfies the query, based on the index's appropriateness score and the item's universal score. Results are shown in order of these probabilities. Appropriateness scores are adjusted based on past user actions (e.g., viewing, purchasing) for similar queries in that index. Appropriateness scores are based on how similar historical queries are to the current query.
10. A computer-implemented method according to claim 1 , wherein: submitting the query to the plurality of search indexes comprises providing a first copy of the query to a first search index and providing a second copy of the query to a second search index; and receiving one of the plurality of search index result sets in response to submitting the query comprises receiving a first search index result set ranked according to a first ranking scale and receiving a second search index result set ranked according to a second ranking scale.
A computer system creates a unified search result from multiple search indexes (e.g., web, products, images). Each index might rank results differently. The system receives a search query and sends it to each index. For each index, it calculates an "appropriateness score" based on how well that index's category suits the query, using past similar queries. It normalizes each item's rank from each index to a common scale to produce a "universal item score." For each item, it calculates the probability that it satisfies the query, based on the index's appropriateness score and the item's universal score. Results are shown in order of these probabilities. Appropriateness scores are adjusted based on past user actions (e.g., viewing, purchasing) for similar queries in that index. The query is sent as potentially different copies to different search indexes. A first search index result set is ranked according to a first ranking scale and a second search index result set is ranked according to a second ranking scale.
11. A computer-implemented method according to claim 10 , wherein determining probabilities of items in the plurality of search index result sets comprises: determining a first appropriateness score for the first search index and determining a second appropriateness score for the second search index; determining a first universal item score for a first item in the first search index result set and determining a second universal item score for a second item in the second search index result set; determining a first probability that the first item satisfies the query based at least in part on the first appropriateness score and the first universal item score; and determining a second probability that the second item satisfies the query based at least in part on the second appropriateness score and the second universal item score.
A computer system creates a unified search result from multiple search indexes (e.g., web, products, images). Each index might rank results differently. The system receives a search query and sends it as potentially different copies to different search indexes. A first search index result set is ranked according to a first ranking scale and a second search index result set is ranked according to a second ranking scale. For each index, it calculates an "appropriateness score" based on how well that index's category suits the query, using past similar queries. It normalizes each item's rank from each index to a common scale to produce a "universal item score." For each item, it calculates the probability that it satisfies the query, based on the index's appropriateness score and the item's universal score. Results are shown in order of these probabilities. Appropriateness scores are adjusted based on past user actions (e.g., viewing, purchasing) for similar queries in that index. A first appropriateness score is determined for the first search index and a second appropriateness score is determined for the second search index; a first universal item score is determined for a first item in the first search index result set and a second universal item score is determined for a second item in the second search index result set; a first probability is determined that the first item satisfies the query based on the first appropriateness score and the first universal item score; and a second probability is determined that the second item satisfies the query based on the second appropriateness score and the second universal item score.
12. A computer system, comprising: at least one processor; memory including instructions that, when executed by the at least one processor, cause the computer system to: receive a query from a searching entity; submit the query to a plurality of search indexes, each search index corresponding to a respective category of items, at least one of the plurality of search indexes utilizing at least one of a different ranking property, scale, function, or definition for ranking items relative to other search indexes; receive one of a plurality of search index result sets from each of the plurality of search indexes in response to the query; determine a plurality of appropriateness scores each corresponding to one of the plurality of search indexes, each of the plurality of appropriateness scores indicating an appropriateness of the category of items corresponding to the respective search index with respect to the query and being based at least in part on historical queries similar to the query that were submitted to the respective search index; determine a universal item score for each of a plurality of items in the plurality of search index result sets at least in part by normalizing the at least one of the different ranking property, scale, function, and definition to a ranking scale common to all the search index result sets; for each of the plurality of items, determine a probability that the item satisfies the query based at least in part on the appropriateness score for the search index associated with the item and the universal item score for the item; include in the universal query result set ones of the plurality of items selected in an order based at least in part on the probabilities of the plurality of items satisfying the query; and provide the universal query result set to the searching entity, wherein determining the plurality of appropriateness scores comprises modifying the plurality of appropriateness scores differently based at least in part on different types of recorded actions associated with the historical queries that were submitted to a corresponding search index.
A computer system creates a unified search result from multiple search indexes (e.g., web, products, images). Each index might rank results differently. The system receives a search query and sends it to each index. For each index, it calculates an "appropriateness score" based on how well that index's category suits the query, using past similar queries. It normalizes each item's rank from each index to a common scale to produce a "universal item score." For each item, it calculates the probability that it satisfies the query, based on the index's appropriateness score and the item's universal score. Results are shown in order of these probabilities. Appropriateness scores are adjusted based on past user actions (e.g., viewing, purchasing) for similar queries in that index. The system contains at least one processor and a memory to store the instructions to perform the calculations.
13. A computer system according to claim 12 , wherein the query is received from a computing device distinct from the computer system comprising the query controller.
A computer system creates a unified search result from multiple search indexes. Each index might rank results differently. The system receives a search query and sends it to each index. The query is received from a separate computing device. For each index, it calculates an "appropriateness score" based on how well that index's category suits the query, using past similar queries. It normalizes each item's rank from each index to a common scale to produce a "universal item score." For each item, it calculates the probability that it satisfies the query, based on the index's appropriateness score and the item's universal score. Results are shown in order of these probabilities. Appropriateness scores are adjusted based on past user actions (e.g., viewing, purchasing) for similar queries in that index.
14. A computer system according to claim 12 , wherein the plurality of search indexes reside, at least in part, on at least one computing device distinct from the computer system receiving the query.
A computer system creates a unified search result from multiple search indexes. Each index might rank results differently. The system receives a search query and sends it to each index. The search indexes reside, at least in part, on one or more computing devices distinct from the computer system receiving the query. For each index, it calculates an "appropriateness score" based on how well that index's category suits the query, using past similar queries. It normalizes each item's rank from each index to a common scale to produce a "universal item score." For each item, it calculates the probability that it satisfies the query, based on the index's appropriateness score and the item's universal score. Results are shown in order of these probabilities. Appropriateness scores are adjusted based on past user actions (e.g., viewing, purchasing) for similar queries in that index.
15. A computer system according to claim 12 , wherein: submitting the query to the plurality of search indexes comprises providing a first copy of the query to a first search index and providing a second copy of the query to a second search index; and receiving one of the plurality of search index result sets in response to submitting the query comprises receiving a first search index result set ranked according to a first ranking scale and receiving a second search index result set ranked according to a second ranking scale.
A computer system creates a unified search result from multiple search indexes (e.g., web, products, images). Each index might rank results differently. The system receives a search query and sends it as potentially different copies to different search indexes. A first search index result set is ranked according to a first ranking scale and a second search index result set is ranked according to a second ranking scale. For each index, it calculates an "appropriateness score" based on how well that index's category suits the query, using past similar queries. It normalizes each item's rank from each index to a common scale to produce a "universal item score." For each item, it calculates the probability that it satisfies the query, based on the index's appropriateness score and the item's universal score. Results are shown in order of these probabilities. Appropriateness scores are adjusted based on past user actions (e.g., viewing, purchasing) for similar queries in that index.
16. A computer system according to claim 15 , wherein determining probabilities of items in the plurality of search index result sets comprises: determining a first appropriateness score for the first search index and determining a second appropriateness score for the second search index; determining a first universal item score for a first item in the first search index result set and determining a second universal item score for a second item in the second search index result set; determining a first probability that the first item satisfies the query based at least in part on the first appropriateness score and the first universal item score; and determining a second probability that the second item satisfies the query based at least in part on the second appropriateness score and the second universal item score.
A computer system creates a unified search result from multiple search indexes (e.g., web, products, images). Each index might rank results differently. The system receives a search query and sends it as potentially different copies to different search indexes. A first search index result set is ranked according to a first ranking scale and a second search index result set is ranked according to a second ranking scale. For each index, it calculates an "appropriateness score" based on how well that index's category suits the query, using past similar queries. It normalizes each item's rank from each index to a common scale to produce a "universal item score." For each item, it calculates the probability that it satisfies the query, based on the index's appropriateness score and the item's universal score. Results are shown in order of these probabilities. Appropriateness scores are adjusted based on past user actions (e.g., viewing, purchasing) for similar queries in that index. A first appropriateness score is determined for the first search index and a second appropriateness score is determined for the second search index; a first universal item score is determined for a first item in the first search index result set and a second universal item score is determined for a second item in the second search index result set; a first probability is determined that the first item satisfies the query based on the first appropriateness score and the first universal item score; and a second probability is determined that the second item satisfies the query based on the second appropriateness score and the second universal item score.
17. A computer system according to claim 12 , wherein at least one of the historical queries is associated with a recorded action of selecting a thereby identified item for viewing.
A computer system creates a unified search result from multiple search indexes (e.g., web, products, images). Each index might rank results differently. The system receives a search query and sends it to each index. For each index, it calculates an "appropriateness score" based on how well that index's category suits the query, using past similar queries. It normalizes each item's rank from each index to a common scale to produce a "universal item score." For each item, it calculates the probability that it satisfies the query, based on the index's appropriateness score and the item's universal score. Results are shown in order of these probabilities. Appropriateness scores are adjusted based on past user actions (e.g., viewing, purchasing) for similar queries in that index. Historical queries are associated with viewing an item.
18. A computer system according to claim 12 , wherein at least one of the historical queries is associated with a recorded action of selecting a thereby identified item for potential purchase.
A computer system creates a unified search result from multiple search indexes (e.g., web, products, images). Each index might rank results differently. The system receives a search query and sends it to each index. For each index, it calculates an "appropriateness score" based on how well that index's category suits the query, using past similar queries. It normalizes each item's rank from each index to a common scale to produce a "universal item score." For each item, it calculates the probability that it satisfies the query, based on the index's appropriateness score and the item's universal score. Results are shown in order of these probabilities. Appropriateness scores are adjusted based on past user actions (e.g., viewing, purchasing) for similar queries in that index. Historical queries are associated with selecting an item for potential purchase.
19. A non-transitory computer-readable storage medium having stored thereon a computer-executable program that when executed directs a computing system to, at least: receive a query from a searching entity; submit the query to a plurality of search indexes, each search index corresponding to a respective category of items, at least one of the plurality of search indexes utilizing at least one of a different ranking property, scale, function, or definition for ranking items relative to other search indexes; receive one of a plurality of search index result sets from each of the plurality of search indexes in response to the query; determine a plurality of appropriateness scores each corresponding to one of the plurality of search indexes, each of the plurality of appropriateness scores indicating an appropriateness of the category of items corresponding to the respective search index with respect to the query and being based at least in part on historical queries similar to the query that were submitted to the respective search index; determine a universal item score for each of a plurality of items in the plurality of search index result sets at least in part by normalizing the at least one of the different ranking property, scale, function, and definition to a ranking scale common to all the search index result sets; for each of the plurality of items, determine a probability that the item satisfies the query based at least in part on the appropriateness score for the search index associated with the item and the universal item score for the item; include in the universal query result set ones of the plurality of items selected in an order based at least in part on the probabilities of the plurality of items satisfying the query; and provide the universal query result set to the searching entity, wherein determining the plurality of appropriateness scores comprises modifying the plurality of appropriateness scores differently based at least in part on different types of recorded actions associated with the historical queries that were submitted to a corresponding search index.
A non-transitory computer-readable medium stores instructions for creating a unified search result from multiple search indexes (e.g., web, products, images). Each index might rank results differently. The system receives a search query and sends it to each index. For each index, it calculates an "appropriateness score" based on how well that index's category suits the query, using past similar queries. It normalizes each item's rank from each index to a common scale to produce a "universal item score." For each item, it calculates the probability that it satisfies the query, based on the index's appropriateness score and the item's universal score. Results are shown in order of these probabilities. Appropriateness scores are adjusted based on past user actions (e.g., viewing, purchasing) for similar queries in that index.
20. A non-transitory computer-readable storage medium according to claim 19 , wherein the probabilities of the items satisfying the query are further based at least in part on allocation scores for the plurality of search indexes, the allocation scores for the plurality of search indexes based at least in part on relative numbers of search results in the plurality of search index result sets generated by the plurality of search indexes.
A non-transitory computer-readable medium stores instructions for creating a unified search result from multiple search indexes. Each index might rank results differently. The system receives a search query and sends it to each index. For each index, it calculates an "appropriateness score" based on how well that index's category suits the query, using past similar queries. It normalizes each item's rank from each index to a common scale to produce a "universal item score." For each item, it calculates the probability that it satisfies the query, based on the index's appropriateness score and the item's universal score. Results are shown in order of these probabilities. Appropriateness scores are adjusted based on past user actions for similar queries. The probabilities of items satisfying the query are further based on allocation scores for the search indexes, based on the relative numbers of search results in each index's result set.
21. A non-transitory computer-readable storage medium according to claim 20 , wherein the probabilities of the items satisfying the query are based on combinations of the plurality of appropriateness scores, the allocation scores for the plurality of search indexes, and the universal item scores for the items.
A non-transitory computer-readable medium stores instructions for creating a unified search result from multiple search indexes. Each index might rank results differently. The system receives a search query and sends it to each index. For each index, it calculates an "appropriateness score" based on how well that index's category suits the query, using past similar queries. It normalizes each item's rank from each index to a common scale to produce a "universal item score." For each item, it calculates the probability that it satisfies the query, based on the index's appropriateness score and the item's universal score. Results are shown in order of these probabilities. Appropriateness scores are adjusted based on past user actions for similar queries. The probabilities of items satisfying the query are further based on allocation scores for the search indexes, based on the relative numbers of search results in each index's result set. The probabilities of the items satisfying the query are based on combinations of the appropriateness scores, the allocation scores, and the universal item scores.
22. A non-transitory computer-readable storage medium according to claim 21 , wherein the combinations comprise linear combinations.
A non-transitory computer-readable medium stores instructions for creating a unified search result from multiple search indexes. Each index might rank results differently. The system receives a search query and sends it to each index. For each index, it calculates an "appropriateness score" based on how well that index's category suits the query, using past similar queries. It normalizes each item's rank from each index to a common scale to produce a "universal item score." For each item, it calculates the probability that it satisfies the query, based on the index's appropriateness score and the item's universal score. Results are shown in order of these probabilities. Appropriateness scores are adjusted based on past user actions for similar queries. The probabilities of items satisfying the query are further based on allocation scores for the search indexes, based on the relative numbers of search results in each index's result set. The probabilities of the items satisfying the query are based on combinations of the appropriateness scores, the allocation scores, and the universal item scores, where the combinations are linear combinations.
23. A non-transitory computer-readable storage medium according to claim 19 , wherein at least one of the historical queries is associated with a recorded action of purchasing a thereby identified item.
A non-transitory computer-readable medium stores instructions for creating a unified search result from multiple search indexes (e.g., web, products, images). Each index might rank results differently. The system receives a search query and sends it to each index. For each index, it calculates an "appropriateness score" based on how well that index's category suits the query, using past similar queries. It normalizes each item's rank from each index to a common scale to produce a "universal item score." For each item, it calculates the probability that it satisfies the query, based on the index's appropriateness score and the item's universal score. Results are shown in order of these probabilities. Appropriateness scores are adjusted based on past user actions (e.g., viewing, purchasing) for similar queries in that index. Historical queries are associated with purchasing an item.
24. A non-transitory computer-readable storage medium according to claim 19 , wherein at least one of the historical queries is associated with a recorded action of selecting a category of a thereby identified item.
A non-transitory computer-readable medium stores instructions for creating a unified search result from multiple search indexes (e.g., web, products, images). Each index might rank results differently. The system receives a search query and sends it to each index. For each index, it calculates an "appropriateness score" based on how well that index's category suits the query, using past similar queries. It normalizes each item's rank from each index to a common scale to produce a "universal item score." For each item, it calculates the probability that it satisfies the query, based on the index's appropriateness score and the item's universal score. Results are shown in order of these probabilities. Appropriateness scores are adjusted based on past user actions (e.g., viewing, purchasing) for similar queries in that index. Historical queries are associated with selecting a category of an item.
25. A non-transitory computer-readable storage medium according to claim 19 , wherein determining the plurality of appropriateness scores comprises determining the plurality of appropriateness scores based at least in part on times associated with the historical queries that were submitted to a corresponding search index.
A non-transitory computer-readable medium stores instructions for creating a unified search result from multiple search indexes (e.g., web, products, images). Each index might rank results differently. The system receives a search query and sends it to each index. For each index, it calculates an "appropriateness score" based on how well that index's category suits the query, using past similar queries. It normalizes each item's rank from each index to a common scale to produce a "universal item score." For each item, it calculates the probability that it satisfies the query, based on the index's appropriateness score and the item's universal score. Results are shown in order of these probabilities. Appropriateness scores are adjusted based on past user actions (e.g., viewing, purchasing) for similar queries in that index. Appropriateness scores are based on the times the historical queries were submitted.
26. A non-transitory computer-readable storage medium according to claim 19 , wherein determining the plurality of appropriateness scores comprises determining the plurality of appropriateness scores based at least in part on how similar the historical queries submitted to a corresponding search index are to the query.
A non-transitory computer-readable medium stores instructions for creating a unified search result from multiple search indexes (e.g., web, products, images). Each index might rank results differently. The system receives a search query and sends it to each index. For each index, it calculates an "appropriateness score" based on how well that index's category suits the query, using past similar queries. It normalizes each item's rank from each index to a common scale to produce a "universal item score." For each item, it calculates the probability that it satisfies the query, based on the index's appropriateness score and the item's universal score. Results are shown in order of these probabilities. Appropriateness scores are adjusted based on past user actions (e.g., viewing, purchasing) for similar queries in that index. Appropriateness scores are based on how similar historical queries are to the current query.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
June 3, 2010
July 2, 2013
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.